Download Microphones

Andrea Mori − matr. 224653 − Lezione del 14/12/2012 − ora 14:30−15.40
Type of microphones
Cables
INTRODUCTION
Preamplifiers
ADC (analog to digital converter)
Mono followed by amplitude panning
Stereo ORTF
TECHNIQUES EMPLOYED
FOR SOUND RECORDING
AND REPRODUCTION
Binaural stereo dipole
Discrete microphone array ITU 5.1
Full ambisonics 3D
Hybrid methods
-1-
Andrea Mori − matr. 224653 − Lezione del 14/12/2012 − ora 14:30−15.40
Microphones
Sound is recorded by microphones which are transducers able to convert
an acoustic signal into an electric one. There are three types of microphones:
 Pressure microphones which are sensitive to acoustic pressure
 Velocity microphones which are sensitive to particle velocity
 Hybrid microphones which are a combination of both quantities
A pressure microphone is completely omnidirectional, it receives sound from
every direction. A velocity microphone is sensitive only to one of the three
Cartesian components of the particle velocity vector, and hence presents a
“figure of eight” directivity pattern, with reversed polarity when the sound
comes form the “negative direction”.
Here a number of different microphone directivity patterns are shown. The
first number is the percentage of sensitivity for pressure and the second for
particle velocity.
Cables
Cables are used to transmit the electric signal. A lot of signal
contaminations can occur inside the cable. To solve this problem you need to
use a balanced cable, with two wires carrying two identical copies of the
signal with opposite polarity; the receiver records the difference between the
voltage carried by the two cables, so that every noise or contamination isn’t
recorded because is the same for each wire.
-2-
Andrea Mori − matr. 224653 − Lezione del 14/12/2012 − ora 14:30−15.40
Preamplifiers
The electric signal arrives to a preamplifiers which not only amplifies the
signal but also processes it. The preamplifier can modify the frequency
response, like an equalizer, but it can also add distortions to create a “better”
musical sound.
ADC (analog to digital converter)
There are two different types of ADC converters:
 PCM converters (like CD, DVD and DAT)
 Bitstream converters (DSD)
In the first type of converters the signal is represented by a series of
integer numbers, with a number of bits between 16 and 24, and a sampling
rate ranging between 44.1, 48, 96 or even 192 kHz.
In bitstream converters the sample frequency is much higher than PCM
converters (typically 2.88 MHz) but there is a binary signal; each sample has
just one bit: 1 or 0.
Bitstream converters are only employed within the Super Audio CD system.
Techniques employed for sound recording and reproductions
Mono followed by amplitude panning
This is a traditional method of recording and reproduction. The recording
is mono; for example in an orchestra each musician has his microphone
which records one track. Than all tracks are mixed with a surround panner so
that it is possible to choose how to distribute the sound among the
loudspeakers.
-3-
Andrea Mori − matr. 224653 − Lezione del 14/12/2012 − ora 14:30−15.40
There are two types of panning:
 Pairwise panning where each signal is always sent to just two
loudspeakers.
 Peter Craven panning where all the loudspeakers are always fed.
ORTF Stereo
This technique is obtained using two microphones and two loudspeakers.
Each microphone feeds the corresponding loudspeakers. Microphones
must be cardioid and they must be separated by a distance of 170-180 mm
and they must create an angle of 110°. Than the best playback configuration
is obtained when two loudspeakers and the listener form an equilateral
triangle. The reproduction of sound is optimal only with this configuration.
Binaural stereo dipole
This method provides an original two channels recording of the signal
coming from N sources and the reproduction occurs over two loudspeakers
which create with the listener an angle of 20°.
-4-
Andrea Mori − matr. 224653 − Lezione del 14/12/2012 − ora 14:30−15.40
Sound is filtered by a digital system called cross talk canceller which can
cancel the cross sound which leaves the right loudspeaker and arrives to the
left ear and vice versa. To achieve this target it is possible to use a matrix
with four filters and with some calculations each loudspeaker produces an
additional sound, which cancels the cross sound of the opposite loudspeaker.
This system has some disadvantages like the difficulty to reproduce the
low frequencies and the presence of a “sweet spot” which is the only one
point where the reproduction is optimal.
It is possible also to create a dual stereo dipole placing two additional
loudspeakers behind the listener, in order to reproduce the sound also in the
rear hemisphere.
-5-
Andrea Mori − matr. 224653 − Lezione del 14/12/2012 − ora 14:30−15.40
Discrete microphone arrays ITU 5.1
This is an advanced technique with five microphones and five
loudspeakers where each microphone feeds the corresponding loudspeaker.
Many different geometrical layouts exist, like
 Williams MMA
 INA−5
 OCT
 Etc…
These methods had a limited success mostly because of the little control
during recording and during post processing and the technician’s work is very
limited. It is also difficult to add to the recording tracks of separately recorded
voices or instruments and also the fixed angle of covering cannot be adjusted
after the recording is done.
-6-
Andrea Mori − matr. 224653 − Lezione del 14/12/2012 − ora 14:30−15.40
Ambisonic 3D
In this method the stage of capturing (or synthesizing) the signal is
separated from the stage of reproduction.
For live recordings, a Soundfield™ microphone must be used, which
captures a B−format 4−channels signal.
This particular signal is composed by four signals, W, X, Y, Z: W
represents the omnidirectional pressure , whilst X, Y and Z represent the
particle velocity in the three Cartesian directions.
The reproduction occurs over an array of 8−24 loudspeakers trough an
Ambisonics decoder.
Today tetrahedral microphone arrays with four cardioid capsules are
usually utilised, which can capture an A−format signal (the 4 “raw” capsule
signals). This is then converted to the standard B−format signal by means of
digital processing, usually performed by software.
In some cases, the B-format signal can be synthesized, starting from one
or more mono sound tracks, which are “ambisonically-panned” by employing
-7-
Andrea Mori − matr. 224653 − Lezione del 14/12/2012 − ora 14:30−15.40
a software which “places” each soundtrack at user’s specified azimuth and
elevation angles. This way of creating a synthetic B-format signal is usually
called “encoding”.
At playback (usually called “decoding” in Ambisonics) the signal which is
sent to each loudspeaker is a weighted mix of the four B−format channels,
with weighting factors which only depends on the position of each
loudspeaker; in fact these factors are the cosines of angles formed by
loudspeakers with the three Cartesian directions.
It is also possible to add a small FIR filter for each loudspeaker, for
matching perfectly the gain and phase of all loudspeakers.
The system shown in the following figure is a three-dimensional
Ambisonics loudspeaker system, and the loudspeakers are placed on two
squares, one vertical and one horizontal.
The signal decoding to feed the loudspeakers is done trough a DSP
platform or with a computer. The DSP has four input channels and initially
signals are split in high and low frequencies, because in the two cases
different decoding coefficients are used, but then all signals are mixed
together. Finally each speaker feed is equalised through an FIR filter.
-8-
Andrea Mori − matr. 224653 − Lezione del 14/12/2012 − ora 14:30−15.40
Hybrid method (Ambiophonics)
This technique combines an Ambisonics 3D system with a stereo dipole
and merges the advantages and disadvantages of each method. For
example the reproduction is optimal only in one point called “sweet spot”
because of the presence of the stereo dipole but the entire system has a
good surround effect due to the Ambisonics 3D system.
-9-